JOURNAL OF BACTERIOLOGY, Aug. 1987, p. 3712-3719 Vol. 169, No. 8 0021-9193/87/083712-08$02.00/0 Copyright © 1987, American Society for Microbiology Cross-Induction of the L- System by L-Rhamnose in Y.-M. CHEN,1 J. F. TOBIN,2 Y. ZHU,' R. F. SCHLEIF,2 AND E. C. C. LIN'* Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, Massachusetts 02115,1 and Department ofBiochemistry, Brandeis University, Waltham, Massachusetts 022542 Received 6 January 1987/Accepted 21 May 1987

Dissimilation of L-fucose as a carbon and energy source by Escherichia coli involves a permease, an isomerase, a kinase, and an aldolase encoded by the fuc regulon at minute 60.2. Utilization of L-rhamnose involves a similar set of proteins encoded by the rha operon at minute 87.7. Both pathways lead to the formation of L-lactaldehyde and phosphate. A common NAD-linked oxidoreductase encoded by fucO serves to reduce L-lactaldehyde to L-1,2-propanediol under anaerobic growth conditions, irrespective of whether the aldehyde is derived from fucose or rhamnose. In this study it was shown that anaerobic growth on rhamnose induces expression of not only thefucO gene but also the entirefuc regulon. Rhamnose is unable to induce the fuc genes in mutants defective in rhaA (encoding L-rhamnose isomerase), rhaB (encoding L-rhamnulose kinase), rhaD (encoding L-rhamnulose 1-phosphate aldolase), rhaR (encoding the positive regulator for the rha structural genes), or fucR (encoding the positive regulator for the fuc regulon). Thus, cross-induction of the L-fucose enzymes by rhamnose requires formation of L-lactaldehyde; either the aldehyde itself or the L- 1-phosphate (known to be an effector) formed from it then interacts with the fucR-encoded protein to induce the fuc regulon.

L-Fucose and L-rhamnose are dissimilated by Escherichia MATERIALS AND METHODS coli in parallel ways (Fig. 1). The trunk pathway for each compound is mediated by a permease (21; J. Power, personal Chemicals. L-Lactaldehyde was prepared by the reaction communication), an isomerase (18, 41, 44), a kinase (12, 23, of ninhydrin with D-threonine (46). L-Rhamnulose was pre- 42, 45), and an aldolase (13, 14, 17, 35, 36). The two pared from a solution of L-rhamnose heated with pyridine differ in the stereoconfiguration at carbons 2 and 4, but (26). L-[U-14C]rhamnose (45 to 60 mCi per milliatom of structural differences in the intermediates disappear with carbon) was purchased from Research Products Interna- cleavage of the phosphorylated by the aldolase, tional Corp., Mount Prospect, Ill. L-Fucose, L-rhamnose, yielding, in both cases, dihydroxyacetone phosphate and and D- were obtained from Sigma Chemical Co., St. L-lactaldehyde. Louis, Mo. Vitamin-free casein acid hydrolysate (CAA) was Aerobically, L-lactaldehyde is converted by an NAD- from ICN Nutritional Biochemicals, Cleveland, Ohio. The linked dehydrogenase to L-lactate, which is oxidized to pyruvate kinase-L-lactate dehydrogenase mixture was from pyruvate for further metabolism (15, 39); anaerobically, Boehringer Mannheim Biochemicals, Indianapolis, Ind. 5- L-lactaldehyde is reduced to L-1,2-propanediol, which is Bromo-4-chloro-3-indolyl-,3-D-galactoside (X-Gal) was from excreted into the medium (15, 40). The sacrifice of the Bachem Inc., Torrance, Calif. All other chemicals were aldehyde as a hydrogen sink increases the portion of commercial products of reagent grade. dihydroxyacetone phosphate that can be utilized as a carbon Bacteria and phage. The E. coli strains used are described and energy source. in Table 1. Eight independent clones of strain ECL116 were The catalytic proteins in each trunk pathway and the mutagenized with ethyl methanesulfonate (28). Following corresponding positive regulatory protein are encoded by a overnight growth of the treated populations in - single gene cluster: thefuc locus at minute 60.2 (1, 7, 16, 37, mineral medium, a portion of each culture was enriched in 38) and the rha locus at minute 87.7 (1, 34). The structural rhamnose-negative mutants by the streptozotocin selection genes of the fuc system appear to be organized as a regulon method (25). Survivors were plated on MacConkey agar comprising at least three operons (7, 20-22), with L-fuculose containing 1% rhamnose as the . Pale-colored colonies 1-phosphate as the effector (3). The rha system responds to were screened on minimal agar containing rhamnose, fu- L-rhamnose as the effector (34). cose, or glucose. Those that failed to grow only on rhamnose A common enzyme of broad function, encoded by the ald were collected and examined for the specific nature of the gene, which is linked to neither the fuc nor the rha locus, is lesion. Deficiencies of L-rhamnose permease, L-rhamnose responsible for the dehydrogenation of L-lactaldehyde to isomerase, and L-rhamnulose kinase activities were detected L-lactate (Y.-M. Chen, Y. Zhu, and E. C. C. Lin, J. by examining cells grown on CAA in the presence of Bacteriol., in press). Likewise, the reduction of L-lactal- rhamnose. Lack of L-rhamnulose 1-phosphate aldolase ac- dehyde to L-1,2-propanediol is catalyzed by a common tivity was diagnosed by inhibition of growth on glycerol in enzyme. However, the gene encoding this enzyme, fucO, is the presence of rhamnose (excessive accumulation of a a member of thefuc regulon (5, 6, 9, 11, 15). How rhamnose phosphorylated sugar causes stasis). Impairment of the causes the induction offucO is the subject of this report. activator protein encoded by rhaC (recently resolved into two genes, rhaR and rhaS [see Results]) was revealed by pleiotropic defects in the enzymes of the rhamnose pathway * Corresponding author. (34) and by complementation with a plasmid, pJTC9, bearing 3712 VOL. 169, 1987 E. COLI L-FUCOSE SYSTEM CROSS-INDUCTION BY L-RHAMNOSE 3713

L-FUCOSE 1- 9:- - :; ¢-. - ::;I: *--.:: ::;:::.-::- 1 :. :; -::--::;:'. -;- v H ~ H f Pemrnwo H*OH HO)AOH L-FUCOSE

~~HNHO-r L-FUCULOSEJ4isomemes OH H OH OH ATPQ L-FUCOSE L-RHAMNS ADPJ;Kinosc L-FUCULOSE 1-PHOSPHATE .i t AkioAose I CH20H CH20H CHO C02 HO-C-H HO-C-H HO-C-H C=O CH3 CH3 CH3 Gee CH2-O-P DIHYDROXYACETONE L-,2-PRPANEDI L-LACTALDEHYDE 11 ---* L-LACTATEA * PHOSPHATE NAD NADH NAD NAWH

A Akdduse

L-RHAMNULISE 1- PHOSPHATE ADP i ATPo L-RHAMNULOSE 11/so L-RHAMNOSE 1.ffnes 70-71. ie 'I ;, 11I,. L-1,2-M .AANEDOL L-RHAMNOSE FIG. 1. Catabolic pathways for L-rhamnose and L-fuCose in E. coli. rhaR+ rhaS+ (see Construction of plasmids, below). Table 2 Strain ECL711 was obtained by transducing the rhaD62 summarizes the characterization of the rhamnose-negative mutation in strain ECL484 (by selecting for a TnJO placed mutants. Strain ECL378 (TnJO 80% linked to rha+) was used nearby) into strain ECL326. as a donor by P1 vir transduction (30) to test whether or not Growth conditions. Quantitative comparisons of enzyme the mutations in rhamnose-negative strains ECL714 (rhaB), activities were carried out with extracts of cells grown ECL715 (rhaA), ECL716 (rhaD), and ECL717 (rhaR) were anaerobically at 37°C in 150-ml flasks filled to the top with at the rha locus. Strain ECL378 was constrtucted by trans- medium, tightly capped, and gently stirred by a magnet. A ducing the TnJO from strain ECL339 (zig-l::TnJO Arha) to mineral solution (43) supplemented with CAA (0.5%), pyru- strain ECL116 (rha+) by selection on LB agar containing vate (30 mM), and thiamine (2 ,ug/ml) was used as the tetracycline and scoring for red colonies on MacConkey- noninducing medium. The inducing medium also contained rhamnose-tetracycline agar. fucose or rhamnose (0.4%). For growth of strains carrying a The following procedure was used to generate an arg plasmid vector, ampicillin (200 ,uag/ml) was added to the deletion extending into the fuc region. The arg::TnJO in medium. For selecting or scoring antibiotic resistance, am- strain ECL357 was transduced into strain ECL56, in which picillin was added to 200 ,ug/ml and tetracycline was added to all of the fucose enzymes were expressed conistitutively. A 20 ,ug/ml. For the preparation of plasmid and chromosomal Tcr Arg- Fuc+ (constitutive) transductant was used for DNA, the cells were grown to stationary phase in LB generatingfuc deletion mutants by selection against the TnJO medium. conferring tetracycline resistance (4). Strain ECL478, iden- Preparation of cell extracts and enzyme assays. Cells har- tified as a Fuc- clone on MacConkey-fucose agar, was found vested from an exponentially growing culture (150 to 200 to lack significant activities of all of the fucose enzymes and Klett units; no. 42 filter) were centrifuged and washed once was thus presumed to have sustained a deletion extending with 0.1 M potassium phosphate (pH 7.0). The final pellet from arg through the entirefuc region. To place a TnJO close was weighed and dispersed in 4 volumes of the same buffer, to the arg-fuc deletion, the following procedure was used. and the cells were disrupted (1 min/ml of suspension) in a Strain ECL289 (eno fuc+ argA::TnlO) was first selected for tube by a model 60 W ultrasonic disintegrator (MSE) while the excision of TnJO (4). A derivative obtained (eno fuc+ being chilled in a - 10°C bath. The extract was centrifuged at AargA7) was used as the recipient for transduction with a P1 100,000 x g for 2 h at 4°C, and the supernatant fraction was lysate prepared from a population with random TnJO inser- used for enzyme assays. tions in the chromosome. A Tcr Eno' transductant was in L-Fucose isomerase and L-rhamnose isomerase activities turn used as the donor of the two markers with the enofuc+ were determined from the initial rate of ketose formation by AargA7 strain as the recipient. A transductant, ECL479, was the cysteine-carbazole method (46). L-Rhamnulose kinase found to have a TnJO 94% linked to eno and 46% linked to activity was determined from the rate of NADH oxidation in argA. Strain ECL479 (zfj3: :TnlO) was used as a transduction the presence of L-rhamnulose, ATP, the pyruvate kinase-L- donor to place the TnWO close to the A(fuc-argA) of strain lactate dehydrogenase mixture (8 ,ug/ml), and phosphoenol- ECL478. The A(fuc-argA) was then transduced into strain pyruvate (12). L-1,2-Propanediol oxidoreductase activity ECL116 by selecting for Tcr and scoring for Fuc- and Arg-. was assayed by the rate of L-lactaldehyde-dependent oxida- Strain ECL366 (A[fuc-argA] zfj3::TnlO) was thus obtained. tion of NADH (11). Protein concentrations in cell extracts 3714 CHEN ET AL. J. BACTERIOL.

TABLE 1. E. (oli strains pJTC40 was constructed by cutting plasmid pJTC9 at the

Source or unique BglII restriction site that lies within the ihaS gene, E. co/i strain Genotype reference filling in the site with the Klenow fragment of DNA poly- merase I, and religating the blunt-ended molecule. This Crookes Wild type ATCC 8739 procedure inserts four base pairs at the restriction site and K-12 ECL1 HfrC phoA8 relAl flmA22 27 T2r (X) generates a termination codon with the rhaS coding se- K-12 ECL56 HfrC phoA8 re/Al tonA22 T2r 21 quence. (X) (Fuc+ and constitutive) Transformation of competent bacteria with plasmids was K-12 ECL289 HfrC eno argA::TnlO relAl 7 carried out by the calcium chloride procedure (31). tonA22 T2r (X) Preparation of DNA. Plasmid DNA was prepared from 1.5 K-12 ECL476 HfrC fucA514 phoA8 relAI T. Chakrabarti ml of LB culture by the alkaline lysis method (31). Chromo- tonA22 T2r (X) somal DNA was prepared from 20 ml of LB culture. The K-12 ECL477 HfrC.fucR501 phoA8 re/Al Y.-M. Chen harvested cells were washed, pelleted, and suspended in 2.0 tonA22 T2r (X) ml of 1.5 M NaCl-100 mM EDTA-20 mM Tris hydrochloride K-12 ECL478 HfrC A(fiu(-argA)I phoA8 relAl This study tonA22 T2r (X) at pH 8.0. The suspension was incubated in the presence of K-12 ECL479 HfrC zJj3::Tn/O AargA7 re/Al This study lysozyme (1 mg/ml) for 30 min at 37°C. After the preparation ton22 T2r (X) was frozen (chilled by dry ice in ethanol) and thawed, 80 pl K-12 ECL484 F+ rhaD62 metBI 34 of 3 M Tris hydrochloride (pH 8.8) and 230 Rl of 10% K-12 ECL116 F AlacUl69 endA hsdR thi 2 N-lauroyl sarcosine were added. The contents were gently K-12 ECL326 F- 41fucOl-/ac::X pl(209)] 11 mixed by stirring with a Pasteur pipette. The mixture was AlacUl69 endA hsdR thi extracted with phenol equilibrated with 10 mM Tris hydro- K-12 ECL333 F 't[rhaFl-/ac::K pl(209)] 10 chloride (pH 8.0) by gentle shaking for 30 min. The super- .lacUl69 endA hsdR thi natant fraction was retrieved, and the extraction procedure K-12 ECL335 F- '4[rhaF2-/ac::X pl(209)] 10 AMac Ul69 endA hsdR thli was repeated twice. The aqueous fraction was then ex- K-12 ECL339 F- A(rha-pfkA)15 zig-l::TnlO 10 tracted three times with ether. Any residual ether was /lacU169 endA hsdR t/li removed under a stream of nitrogen. The sample was heated K-12 ECL357 F arg::Tn/O recB21 thyA36 M. Syvanen to 60°C for 10 min and dialyzed overnight against two K-12 ECL366 F A(fuc-argA)IX zfj3::TnlO This study changes of 0.1 mM EDTA-10 mM NaCI-10 mM Tris hydro- AlacUl69 endA hsdR thi chloride at pH 8.0. K-12 ECL378 F zig-l::TnlO A/acU169 endA This study Southern transfer and hybridization. DNA fragments were hsdR thi labeled by nick translation (31). Chromosomal DNA (10 jig) K-12 ECL711 F tL4fucO-/ac::X pl(209)] This study was digested with 100 U of EcoRI, precipitated with ethanol, rhaD62 zig-l::TnlO AlaacU169 endA hsdR thi and electrophoresed on a 0.8% agarose gel. Southern trans- K-12 ECL714 F- rhaB101 AlacUl69 endA This study fers were then performed (31) with T-500 added to h/sdR thi the hybridization buffer at 0.1 g/ml. K-12 ECL715 F- rhaA502 z\/acUl69 endA This study hsdR thi RESULTS K-12 ECL716 F- rhaD701 /lac U169 endA This study hsdR thi Anaerobic inducing effects of fucose and rhamnose. Be- K-12 ECL717 F- rhaR702 Mac Ul69 endA This study cause the fuc structural genes are partitioned in three oper- hsdR thi ons (fiWPIK, encoding L-fucose permease, L-fucose isomerase, and L-fuculose kinase, respectively;fucA, encod- ing L-fuculose 1-phosphate aldolase; and fiucO, encoding were estimated with bovine serum albumin as a standard L-1,2-propanediol oxidoreductase [11, 21, 22; Chen et al., (29). Specific activities of the isomerase, the kinase, and the submitted], cells of a wild-type K-12 Hfr strain (ECL1) were oxidoreductase are expressed in nanomoles per minute per grown anaerobically on CAA and pyruvate, with or without milligram of protein at 25°C. rhamnose, to discover whether cross-induction of the fucose For determination of L-rhamnose permease activity, cells system was limited to L-1,2-propanediol oxidoreductase. grown aerobically in mineral medium supplemented with Enzyme assays of the cell extracts revealed that rhamnose 0.2% rhamnose and 0.5% CAA were incubated for 1 min in cross-induced not only the oxidoreductase but also the a medium containing 20 ,uM labeled substrate. The radioac- tivity retained by the cells, washed on a filter disk, was TABLE 2. Characterization of rhamnose-negative mutants determined (8). The rate of accumulation of radioactivity was linear for 2 min. Permease activity is expressed in Enzyme sp act' Growth nanomoles of rhamnose uptake per minute per milligram E. coli strain response (dry weight) of cells at 25°C. and genotype L-Rhamnose L-Rhamnose L-Rhamnulose rham- Construction of plasmids. Standard molecular cloning pro- permease isomerase kinase noseb cedures were followed (31). The plasmid pfuc20 was con- ECL714 rhaB 11 0 R structed by inserting into a 3,600 pBR322 3.4-kilobase (kb) ECL715 rhaA 11 0 570 R BamHI-EcoRI fragment that contained fii(R but no intact ECL716 rhaD 5 2,600 320 S structural genes of thefuc regulon (Y.-M. Chen, Y. Zhu, and ECL717 rhaR 0.2 5 0 R E. C. C. Lin, submitted for publication). Plasmid pJTC9 was constructed by inserting the 2.2-kb BamHI-to-EcoRI frag- " The cells were grown aerobically in mineral medium containing 0.2% ment of plasmid pJTC5 (J. F. Tobin and R. F. J. rhamnose and 0.5% CAA. Schleif, I Growth was tested on agar containing 0.2% glycerol with or without 0.2% Mol. Biol., in press), a plasmid containing the rhamnose rhamnose. Abbreviations: R, resistant to rhamnose as indicated by normal operons, into the polylinker of plasmid pGC2 (32). Plasmid colony sizes: S. sensitive to rhamnose as indicated by subnormal colony sizes. VOL. 169, 1987 E. COLI L-FUCOSE SYSTEM CROSS-INDUCTION BY L-RHAMNOSE 3715

TABLE 3. Anaerobic induction of the isomerases and L-1,2-propanediol oxidoreductase by fucose and rhamnose in various strains and mutants Enzyme sp act E. coli strain and Inducer added to relevant genotype growth medium L-Rhamnose L-Fucose L-1,2-Propanediol isomerase isomerase oxidoreductase Crookesfuc+ rha+ None 50 35 70 Fucose 200 3,000 1,300 Rhamnose 12,000 3,500 1,200 ECLlfuc' rha+ None 30 20 40 Fucose 20 2,000 950 Rhamnose 7,500 1,900 900 ECL116fuc' rha+ None 20 20 40 Fucose 30 1,500 900 Rhamnose 6,000 1,200 500 ECL339 A(rha-pflcA) Fucose 30 1,500 850 Rhamnose 230 25 40 ECL366 A(fuc-argA) Fucose 20 20 40 Rhamnose 5,000 40 34 permease, the isomerase, the kinase, and the aldolase (data of fucose, rhamnose, L-lactaldehyde, or D-xylose. After not shown). aerobic incubation for about 36 h at 37°C, the colonies that An examination of two other E. coli stocks, a K-12 F- formed around the disk charged with fucose, rhamnose, or strain (ECL116) and strain Crookes (19), indicated that the L-lactaldehyde were blue. In contrast, the colonies on the gratuitous induction of the fucose trunk pathway by plate with D-xylose were uniformly colorless (data not rhamnose was not an idiosyncratic property of the K-12 Hfr shown). In a control experiment with strain ECL711, a strain used. In all three strains, rhamnose induced not only transductant of strain ECL326 which inherited a rhaD mu- L-rhamnose isomerase but also L-fuCOse isomerase and tation abolishing L-rhamnulose 1-phosphate aldolase activ- L-1,2-propanediol oxidoreductase (Table 3). In contrast, ity, ,B-galactosidase was induced by fucose and L-lactal- fucose induced L-fucose isomerase and L-1,2-propanediol dehyde but not by rhamnose. oxidoreductase but not L-rhamnose isomerase. The levels of L-Lactaldehyde may act directly as an effector for induc- the fucose enzymes induced by fucose or rhamnose were tion of thefuc regulon, or it may induce indirectly by giving similar. rise to L-fuculose 1-phosphate through the reversible reac- Chemical contamination of the rhamnose stock by fucose tion catalyzed by the aldolase encoded by fucA. The equi- was ruled out by the inability of the same rhamnose prepa- librium constant for [L-lactaldehyde][dihydroxyacetone ration to induce L-fucose isomerase and L-1,2-propanediol phosphate]/[L-fuculose 1-phosphate] is 0.46 mM (17), and oxidoreductase in strain ECL339 with a rha deletion. Fortu- the intracellular concentration of dihydroxyacetone phos- itous activity of L-rhamnose isomerase on fucose was ex- phate in E. coli is generally about 0.2 mM (27). Conse- cluded by the failure of rhamnose to induce any fucose- quently, if the activity of L-fuculose 1-phosphate aldolase is isomerizing activity in strain ECL366 with a fuc deletion. not limiting, the ratio of L-fuculose-1-phosphate/L-lactalde- Thus, either rhamnose itself or one of its metabolites cross- hyde can approach 0.4. To determine whether L-lactalde- induces the fuc regulon. hyde or L-fuculose 1-phosphate is responsible for induction A derivative of rhamnose as the inducer of thefuc regulon. of the fuc regulon by rhamnose, we tested a fucA point Four rha mutants, defective in different genes, were ana- mutant, ECL476 (<1% aldolase activity when grown under lyzed for induction of the fucose enzymes by rhamnose. A inducing conditions). In this mutant, rhamnose was still able defect in the rhaA gene, encoding L-rhamnose isomerase, to induce L-fucose isomerase and L-1,2-propanediol oxido- prevented induction of L-fucose isomerase and L-1,2- reductase (Table 4). Although the result suggests that L- propanediol oxidoreductase by rhamnose. Thus, rhamnlose lactaldehyde serves as an alternative effector for the fuc itself is incapable of inducing the fuc regulon. The ability of regulon, a definitive conciusion cannot be made until it can rhamnose to induce L-fucose isomerase and L-1,2- be established that the fucA mutation in the test strain is propanediol oxidoreductase was also abolished by a muta- nonleaky and that no significant activity is contributed by tion in rhaB, encoding the kinase; rhaD, encoding the other enzymes in the cell. A small residual enzyme activity aldolase; or rhaR, encoding the activator protein (Table 4). in vivo would suffice to build up an inducing concentration of Rhamnose, therefore, has to be metabolized to or beyond L-fuculose 1-phosphate in the metabolic cul de sac. L-lactaldehyde to induce the fuc genes. As expected, two fucose-negative mutants revealed to be Because the supply of L-lactaldehyde was limited, a small- defective in fucI by complementation with the cloned gene scale qualitative test for its inductive effect was carried out were induced in its fucose permease (fuculose kinase activity with strain ECL326, bearing N(fucO-IacZ). About 103 cells could not be readily assayed because rhamnulose kinase also were spread uniformly on four agar plates containing CAA acted on fuculose [12]). Likewise, two fucose-negative mu- (200 pLg/ml) and X-Gal (40 ,ug/ml), a chromogenic substrate tants revealed to be defective in fucK by complementation of ,B-galactosidase. A sterile filter disk was then placed on with the cloned gene were induced in its fucose permease the center of each agar plate and impregnated with 10 ,umol and fucose isomerase by rhamnose (data not shown). 3716 CHEN ET AL. J. BACTERIOL.

TABLE 4. Activities of the isomerases and L-1,2-propanediol oxidoreductase in rha and fiwc: mutants grown anaerobically Enzyme sp act (U) Strain (plasmid) Rhamnose in and relevant genotype growth medium L-Rhamnose L-Fucose L-1,2-Propanediol isomerase isomerase oxidoreductase ECL116 rha+ fuc 30 20 40 + 6,000 1,200 500 ECL715 rhaA502 10 10 20 + 10 10 15 ECL714 rhaB101 20 20 30 + 3,600 20 40 ECL716 rhaD701 20 110 40 + 2,000 180 50 ECL717 rhaR702 0 0 20 + 5 0 20 ECL476 fucA514 10 10 50 + 6,900 1,700 290" ECL477 fucRS01 40 10 60 + 9,800 10 90 ECL477fucR501 (pfuc20fucR+) - 2 40 70 + 9,600 2,900 1,400 aAn activity of 180 U was observed in cells induced with fucose. The subnormal level might be attributable to divergent but overlapping promoter regions of fucA and fucO (Chen and Zhu, unpublished data).

The role of thefucR gene in the induction of thefuc regulon kb, the corresponding fragments from the P(rhaF-lac) mu- by rhamnose. The possibility of L-lactaldehyde acting as an tants were only 5.4 kb (Fig. 3). Hence, the lac fusion alternative inducer of the fuc regulon in turn raises the introduced a second EcoRI site in the rha region. Since there question of whether an independent regulator gene exists for is only a single EcoRI site in the entire rha region close to the control of thefuc regulon by rhamnose. If so, inactivation of promoter of the rhaS and only a single EcoRI site in the fucR should not abolish the cross-inducing activity of coding region of the lacZ gene, the shortened chromosomal rhamnose. When strain ECL477, afucR mutant, was grown fragment should be the product of cuts at the promoter of anaerobically in the presence of rhamnose, neither L-fucose rhaS and the site within lacZ. The rhaS and rhaR genes span isomerase nor L-1,2-propanediol oxidoreductase was in- about 1.8 kb. The distance from the fusion joint to the EcoRI duced (Table 4). The lost ability was restored by transfor- site in lacZ is deduced to be about 4 kb. The lacZ in mation with a plasmid (pfuc20) bearing the fucR' gene but P(rhaF-lacZ'), therefore, should be fused to the end of the no intact fuc structural genes. It therefore seems that gene rhaR. Furthermore, when plasmid pJTC9, bearing only rhamnose induces the fucose system by virtue of being a rhaR and rhaS, was introduced into strains ECL717 precursor of an effector that interacts with the fucR product. (rhaR702) and ECL333 (4[rhaF-lac]), the structural genes of The rhaF locus is rhaR. It was reported that disruption of both the rha andfuc systems became inducible by rhamnose a gene in the rha region by a lac fusion prevented induction (Table 5). The plasmid pJTC40, bearing rhaR without a of fucO by rhamnose. The suggestion was made that the functional rhaS, was also effective in complementing the fusion occurred in a gene, rhaF, that encoded an activator of defect in the two strains. Hence, the lac fusion was indeed fucO (10). A subsequent finding that the fusion reduced the within the rhaR gene, and the disruption was near its 3' end. inducible level of L-rhamnose isomerase, together with the Subnormal activation of the rha operon would result in results on cross-induction described in the present study, reduced metabolic flow through the rhamnose pathway, raised the possibility that the fusion affected the activator lowering the steady-state concentration of L-lactaldehyde. It gene for the rha structural genes. might be noted that the presence of multicopies of both rhaR The rhaC locus, previously thought to encode the activa- and rhaS resulted in hyperinducibility of L-rhamnose tor protein for the rha system (34), has recently been found isomerase but not that of L-fucose isomerase and L-1,2- to contain two genes: rhaR and rhaS. The protein encoded propanediol oxidoreductase, the induction of which is evi- by rhaR is sufficient for activating the rhaBAD operon dently limited by the single copy of fucR. The presence of (Tobin and Schleif, in press). The location of 'P(rhaF-lacZ) multicopies of rhaR alone did not result in hyperinducibility was first determined by a Southern transfer experiment. Chromosomal DNAs prepared from the wild-type strain rhaS (ECL116) and the two mutants (ECL333 and ECL335) rhaR bearing independent lac fusions were digested with EcoRI. A 0.9-kb fragment of the rhaS gene produced by EcoRI and EcoRI Clal BstEH SmaI BstEII cuts (Fig. 2) was used as a probe to hybridize with the 11 EcoRI-digested chromosomal DNA of each strain. Whereas 0 0.5 1.0 1 5 20kb the hybridizing fragment from the wild-type strain was 6.8 FIG. 2. Restriction map of rhaR and rhaS genes. VOL. 169, 1987 E. COLI L-FUCOSE SYSTEM CROSS-INDUCTION BY L-RHAMNOSE 3717

1 2 3 enzyme activity is useful for hydrogen disposal during an- aerobic growth but is wasteful of carbon and energy source during aerobic growth, a mechanism is necessary for regu- lation of catalytic activity according to the cellular respira- tory state. This is accomplished by posttranslational modi- fication, first shown by the presence of a low specific activity of the enzyme but undiminished levels of the immunochemi- cally cross-reacting material in cells grown aerobically on fucose. Cells grown aerobically on rhamnose, however, were observed not to be induced in this protein (5, 6, 9, 11). Aerobically, rhamnose probably failed to cross-induceffucO because the metabolic flow rate through the trunk pathway was inadequate for maintaining the necesssary L-lactal- dehyde level. The aerobic doubling time on rhamnose was 200 min, in contrast to the 80-min doubling time supported by fucose (Y. Zhu, unpublished data). 6.8-_U A previous study using the technique of lac fusion impli- 5.4- _ 3 cated a regulatory gene in the rha locus for anaerobic cross-induction offucO. In two strains bearing an indepen- dently formed t?(rhaF-1ac), anaerobic induction offucO by FIG. 3. Autoradiogram of Southern blot hybridization between rhamnose no longer occurred (10). The results from South- the 32P-labeled EcoRI-BstEII fragment from the rhaS gene and ern blot hybridizations and the complementation experi- EcoRI restriction digests of chromosomal DNA. Numbers on the ments in the present study revealed that the lac segment, left indicate the sizes of the hybridized chromosomal fragments in instead of being fused to a distinct regulatory gene, rhaF, kilobases. Lanes: 1, DNA from strain ECL116; 2, DNA from strain was joined to the 3' end of rhaR with only partial impairment ECL333; 3, DNA from strain ECL335. of its function. As a consequence, rhamnose could no longer adequately induce its trunk pathway. The more severe reduction in the ability of rhamnose to induce fucO (>10- of L-rhamnose isomerase. It is not certain whether this is fold) than in its ability to induce fucI (<4-fold) and rhaA because the two genes products have a synergistic effect or (<4-fold) can only be tentatively explained. It is possible because in the cells harboring pJTC40 the function of the that the subnormal induction of L-1,2-propanediol oxidore- rhaR gene product was interfered with by an incomplete ductase still gave sufficient enzyme activity to meet the need product of the rhaS gene. It might also be noted that of a lowered metabolic flow rate through the rhamnose trunk ,B-galactosidase was induced by rhamnose in strain ECL333 pathway and that, when the effector-activator complex was (QF[rhaF-lac]) (10; Table 5). These induction patterns indi- limiting, fucI was expressed at a higher level than fucO cated that the regulators of the rha system are under positive because of a difference in promoter affinity. autogenous regulation. This interpretation is consistent with The collective phenotypes of all of the mutations stud- the results of Si mapping experiments examining transcrip- ied-rhaA, rhaB, rhaD, rhaR, andfucR-are consistent with tion from the rhaS and rhaR genes (Tobin and Schleif, in the view that rhamnose induces thefuc system by giving rise press). to L-lactaldehyde (or L-fuculose 1-phosphate). Induction of DISCUSSION the fucose system by rhamnose was previously noted in Klebsiella pneumoniae, but the cause was unexplored (E. J. Aside from being shared and jointly regulated in its St. Martin, Ph.D. thesis, University of Massachusetts, synthesis by fucose and rhamnose, L-1,2-propanediol oxido- Amherst, 1975). The simplest model is that the effector reductase has a more unusual regulatory feature. Since the produced from rhamnose combines with the fucR gene

TABLE 5. Complementation of 4(rhaF-lacZ) in strain ECL333 by rhaR+ rha genotype Rhamnose Enzyme sp acta Strain (plasmid) Host Plasmid in groh L-Rhamnose L-Fucose L-1,2-Propanediol HostPdmedium isomerase isomerase oxidoreductase P-Galactosidase ECL717 rhaR702 0 0 22 + 5 0 20 ECL717(pJTC9) rhaR702 rhaR+ rhaS+ - 20 10 50 + 18,000 1,600 1,000 ECL333 (1(rhaF-lacZ) 0 10 15 15 + 2,000 450 85 250 ECL333(pJTC9) 4(rhaF-lacZ) rhaR+ rhaS+ - 0 20 27 35 + 17,000 1,600 990 1,000 ECL333(pJTC40) '1(rhaF-lacZ) rhaR+ rhaS- - 0 30 17 30 + 7,000 1,500 800 1,000 a All cultures were grown anaerobically with rhamnose as the inducer. 3718 CHEN ET AL. J. BACTERIOL. product to turn on the fiwc regulon. A more complex model properties of L-rhamulokinase. Biochim. Biophys. Acta 92: would be that the cross-induction involves L-lactaldehyde as 489-497. the effector instead of L-fuculose 1-phosphate and that 13. Chiu, T. H., and D. S. Feingold. 1965. Substrate specificity of induction by L-lactaldehyde requires physical interaction L-rhamnulose 1-phosphate aldolase. Biochem. Biophys. Res. Commun. 19:511-516. between the regulatory proteins of the fuc and rha systems. 14. Chiu, T. H., and D. S. Feingold. 1969. L-Rhamnulose 1- Indirect evidence for interaction between the regulatory phosphate aldolase from Escherichia coli. Crystallization and proteins was provided by the observations that mutations in properties. Biochemistry 8:98-108. fucR which render the fiuc regulon inducible by D- 15. Cocks, G. T., J. Aguilar, and E. C. C. Lin. 1974. Evolution of (24) were sometimes accompanied by defects in rhamnose L-1,2-propanediol catabolism in Escherichia coli by recruitment utilization (St. Martin, Ph.D. thesis) and that selection for of enzymes for L-fucose and L-lactate metabolism. J. Bacteriol. growth on the rare sugar L-, an analog of L-fucose, 118:83-88. resulted in loss of the ability to grow on rhamnose (Zhu and 16. Falkow, S., H. Schneider, L. S. Baron, and S. B. Formal. 1963. Lin, manuscript in preparation). Virulence of Escherichia-Shigella genetic hybrids for the guinea pig. J. Bacteriol. 86:1251-1258. The advantage of cross-induction of the fucose pathway 17. Ghalambor, M. A., and E. C. Heath. 1962. The metabolism of by rhamnose through a common metabolite or effector is L-fucose. 11. The enzymatic cleavage of L-fuculose-1-phos- mechanistic simplicity. Such a mechanism, however, exacts phate. J. Biol. Chem. 237:2427-2433. a price of gratuitous synthesis of the enzymes in the fucose 18. Green, M., and S. S. Cohen. 1956. The enzymatic conversion of trunk pathway when only rhamnose is present. The absence L-fucose to L-fuculose. J. Biol. Chem. 219:557-568. of a control by which rhamnose can selectively activatefucO 19. Gunsalus, I. C., and D. B. Hand. 1941. The use of bacteria in the might indicate that during the evolutionary history of E. cobi chemical determination of total vitamin C. J. Biol. Chem. rhamnose was seldom present without fucose. Since both 141:853-858. sugars are frequent constituents of (33), 20. Hacking, A. J., J. Aguilar, and E. C. C. Lin. 1978. Evolution of synthesis of the fucose enzymes propanediol utilization in Escherichia coli: mutant with im- in the trunk pathway might proved substrate-scavenging power. J. Bacteriol. 136:522- rarely be gratuitous. Alternatively, the lack of a more elegant 530. mechanism for the cross-induction indicates that the rha and 21. Hacking, A. J., and E. C. C. Lin. 1976. Disruption of the fucose fuc systems are in the process of progressive, or retrogres- pathway as a consequence of genetic adaptation to propanediol sive, evolution. as a carbon source in Escherichia coli. J. Bacteriol. 126:1166- 1172. ACKNOWLEDGMENT 22. Hacking, A. J., and E. C. C. Lin. 1977. Regulatory changes in the fucose system associated with the evolution of a catabolic This work was supported by grant DMB8314312 from the Na- pathway for propanediol in Escherichia coli. J. Bacteriol. tional Science Foundation. 130:832-838. 23. Heath, E. C., and M. A. Ghalambor. 1962. The metabolism of LITERATURE CITED L-fucose. I. The purification and properties of L-fuculose kinase. 1. Bachmann, B. J. 1983. Linkage map of Escherichia coli K-12, J. Biol. Chem. 237:2423-2426. edition 7. Microbiol. Rev. 47:180-230. 24. LeBlanc, D. J., and R. P. Mortlock. 1971. Metabolism of 2. Backman, K., Y.-M. Chen, and B. Magasanik. 1981. Physical D-arabinose: a new pathway in Escherichia (oli. J. Bacteriol. and genetic characterization of the ginA-ginG region of the 106:90-96. Escherichia c0oli chromosome. Proc. Natl. Acad. Sci. USA 25. Lengeler, J. 1979. Streptozotocin, an antibiotic superior to 78:3743-3747. penicillin in the selection of rare bacterial mutations. FEMS 3. Bartkus, J. M., and R. P. Mortlock. 1986. Isolation of a mutation Microbiol. Lett. 5:417-419. resulting in constitutive synthesis of L-fucose catabolic en- 26. Levene, P. A., and D. W. Hill. 1933. The action of pyridine on zymes. J. Bacteriol. 165:710-714. sugars. J. Biol. Chem. 102:563-571. 4. Bochner, B. R., H.-C. Huang, G. L. Schieven, and B. N. Ames. 27. Lin, E. C. C. 1976. Glycerol dissimilation and its regulation in 1980. Positive selection for loss of tetracycline resistance. J. bacteria. Annu. Rev. Microbiol. 30:535-578. Bacteriol. 143:926-933. 28. Lin, E. C. C., S. A. Lerner, and S. E. Jorgensen. 1962. A method 5. Boronat, A., and J. Aguilar. 1979. Rhamnose-induced for isolating constitutive mutants for -catabolizing propanediol oxidoreductase in Escherichia coli: purification, enzyme. Biochim. Biophys. Acta 60:422-424. properties, and comparison with the fucose-induced enzyme. J. 29. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. G. Randall. Bacteriol. 140:320-326. 1951. Protein measurement with the Folin phenol reagent. J. 6. Boronat, A., and J. Aguilar. 1981. Metabolism of L-fucose and Biol. Chem. 193:265-275. L-rhamnose in Escherichia co/i: differences in induction of 30. Luria, S. E., J. N. Adams, and R. C. Ting. 1960. Transduction of propanediol oxidoreductase. J. Bacteriol. 147:181-185. -utilizing ability among strains of E. coli and S. dysen- 7. Chakrabarti, T., Y.-M. Chen, and E. C. C. Lin. 1984. Clustering teriace and the properties of the transducing phage particles. of genes for L-fucose dissimilation by Escherichia (coi. J. Virology 12:348-390. Bacteriol. 157:984-986. 31. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular 8. Chen, Y.-M., T. Chakrabarti, and E. C. C. Lin. 1984. Consti- cloning: a laboratory manual. Cold Spring Harbor Laboratory, tutive activation of L-fucose genes by an unlinked mutation in Cold Spring Habor, N.Y. Escherichia coli. J. Bacteriol. 159:725-729. 32. Myers, R. M., L. S. Lerman, and T. Maniatis. 1985. A general 9. Chen, Y.-M., and E. C. C. Lin. 1984. Post-transcriptional method for saturation mutagenesis of cloned DNA fragments. control of L-1,2-propanediol oxidoreductase in the L-fucose Science 229:242-247. pathway of Esc/heric/hia (oli K-12. J. Bacteriol. 157:341-344. 33. Pigman, W. (ed). 1957. The : chemistry, biochem- 10. Chen, Y.-M., and E. C. C. Lin. 1984. Dual control of a common istry, and physiology. Academic Press, Inc., New York. L-1,2-propanediol oxidoreductase by L-fucose and L-rhamnose 34. Power, J. 1967. The L-rhamnose genetic system in Escherichia in Escherichia co/i. J. Bacteriol. 157:828-832. (oli K-12. Genetics 55:557-568. 11. Chen, Y.-M., E. C. C. Lin, J. Ros, and J. Aguilar. 1983. Use of 35. Sawada, H., and Y. Takagi. 1964. The metabolism of t- operon fusions to examine the regulation of the L-1,2- rhamnose in Escherichia coli. IlI. L-Rhamnulose-phosphate propanediol oxidoreductase gene of the fucose system in Esc/h- aldolase. Biochim. Biophys. Acta 92:26-32. erichia coli K12. J. Gen. Microbiol. 129:3355-3362. 36. Schwartz, N. B., and D. S. Feingold. 1972. L-Rhamnulose 12. Chiu, T. H., and D. S. Feingold. 1964. The purification and 1-phosphate aldolase from Escherichia co/i. III. The role of VOL. 169, 1987 E. COLI L-FUCOSE SYSTEM CROSS-INDUCTION BY L-RHAMNOSE 3719

divalent cations in enzyme activity. Bioinorg. Chem. 2:75-86. Biophys. Acta 92:10-17. 37. Skjold, A. C., and D. H. Ezekiel. 1982. Analysis of lambda 42. Takagi, Y., and H. Sawada. 1964. The metabolism of L- insertions in the fucose utilization region of Escherichia coli rhamnose in Escherichia coli. II. L-Rhamnulose kinase. Bio- K-12: use of fuc and A argA transducing bacteriophages to chim. Biophys. Acta 92:18-25. partially order the fucose utilization genes. J. Bacteriol. 43. Tanaka, S., S. A. Lerner, and E. C. C. Lin. 1967. Replacement 152:120-125. of a phosphoenolpyruvate-dependent phosphotransferase by 38. Skjold, A. C., and D. H. Ezekiel. 1982. Regulation of D- nicotinamide adenine dinucleotide-linked dehydrogenase for the arabinose utilization in Escherichia coli K-12. J. Bacteriol. utilization of mannitol. J. Bacteriol. 93:642-648. 152:521-523. 44. Wilson, D. M., and S. AjI. 1957. Metabolism of L-rhamnose by 39. Sridhara, S., and T. T. Wu. 1969. Purification and properties of Escherichia coli. I. L-Rhamnose isomerase. J. Bacteriol. lactaldehyde dehydrogenase in Escherichia coli. J. Biol. Chem. 73:410-414. 244:5233-5238. 45. Wilson, D. M., and S. Ajl. 1957. Metabolism of L-rhamnose by 40. Sridhara, S., T. T. Wu, T. M. Chused, and E. C. C. Lin. 1969. Escherichia coli. II. The phosphorylation of L-rhamnulose. J. Ferrous-activated nicotinamide adenine dinucleotide-linked de- Bacteriol. 73:415-420. hydrogenase from a mutant of Escherichia coli capable of 46. Zagalak, B., P. A. Frey, G. L. Karabatsos, and R. M. Abeles. growth on 1,2-propanediol. J. Bacteriol. 98:87-95. 1966. The stereochemistry of the conversion of D- and L-1,2- 41. Takagi, Y., and H. Sawada. 1964. The metabolism of L-rham- propanediols to propionaldehyde. J. Biol. Chem. 241:3028- nose in Escherichia coli. I. L-Rhamnose isomerase. Biochim. 3035.